Exploration strategies for learning in multi-agent foraging

Authors:
Yogeswaran Mohan;Ponnambalam S.G.
Affiliations:
School of Engineering, Monash University, Petaling Jaya, Selangor, Malaysia;School of Engineering, Monash University, Petaling Jaya, Selangor, Malaysia
Venue:
SEMCCO'11 Proceedings of the Second international conference on Swarm, Evolutionary, and Memetic Computing - Volume Part II
Year:
2011

Citing 8
Cited 0

Introduction to Reinforcement Learning

Introduction to Reinforcement Learning
Exploration Strategies for Model-based Learning in Multi-agent Systems: Exploration Strategies

Autonomous Agents and Multi-Agent Systems
PAC model-free reinforcement learning

ICML '06 Proceedings of the 23rd international conference on Machine learning
Empirical Studies in Action Selection with Reinforcement Learning

Adaptive Behavior - Animals, Animats, Software Agents, Robots, Adaptive Systems
The many faces of optimism: a unifying approach

Proceedings of the 25th international conference on Machine learning
Accelerating reinforcement learning through implicit imitation

Journal of Artificial Intelligence Research
Emergence of flocking behavior based on reinforcement learning

KES'06 Proceedings of the 10th international conference on Knowledge-Based Intelligent Information and Engineering Systems - Volume Part III
A new Q-learning algorithm based on the metropolis criterion

IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics

Quantified Score

Hi-index	0.00

Visualization

Abstract

During the learning process, every agent's action affects the interaction with the environment based on the agent's current knowledge and future knowledge. The agent must therefore have to choose between exploiting its current knowledge or exploring other alternatives to improve its knowledge for better decisions in the future. This paper presents critical analysis on a number of exploration strategies reported in the open literatures. Exploration strategies namely random search, greedy, ε-greedy, Boltzmann Distribution (BD), Simulated Annealing (SA), Probability Matching (PM) and Optimistic Initial Values (OIV) are implemented to study on their performances on a multi-agent foraging task modeled.